Embodiments of the present disclosure relate to optimizing neurosynaptic networks, and more specifically, to reducing the number of neurons and axons in a neurosynaptic network while maintaining its functionality.